|
Guide to the Application of Genotyping to Tuberculosis Prevention
and Control
Return to Genotyping Main Menu
Combining Genotyping and Epidemiologic
Data to Improve Our Understanding of
Tuberculosis Transmission
Comparing Genotyping and Epidemiologic Results from Two TB Patients
In this section, we will analyze the most basic situation – the
results of two TB patients. A conceptual framework for understanding
the possible combinations of these results is provided in Table
4.5, which is separated into two categories: a) patients that were
in fact involved in the same chain of recent transmission and b)
patients that were not.
Genotypes Match and Known Epidemiologic Link Identified
If an investigation establishes that two persons with TB share
a known epidemiologic link and their isolates have a matching PCR
genotype, this provides good evidence that the two persons were
involved in the same chain of recent transmission. This often means
that one of the persons transmitted TB to the other, but it is also
possible that both became infected by a third person. Other more
complex transmission scenarios are also possible. Since the evidence
strongly supports recent transmission, no additional information
needs to be collected from a cluster investigation, and an RFLP
analysis is not indicated.
Genotypes Match but No Epidemiologic Link or Only Possible
Epidemiologic Link Identified
In contrast to the previous example, whether two persons are involved
in the same chain of recent transmission is less clear if they have
isolates with matching genotypes but routine case and contact investigations
failed to identify a known link between them. If the two persons
were, in fact, involved in the same chain of recent transmission,
failure to identify when and where transmission occurred could be
the result of investigators not asking the right questions or the
persons interviewed not knowing or being unwilling to give complete
answers. For example, if an outbreak occurs among drug users, it
might take a lengthy investigation conducted by highly skilled interviewers
to discover even a few of the actual epidemiologic links.
Another example of routine investigations being unable to discover
hidden epidemiologic links are instances of casual contact. Casual
contact was once thought to be a rare cause of TB transmission,
but recent epidemiologic studies relying on genotyping results have
shown that this concept needs to be updated. In a study conducted
by the Maryland TB program, intensive interviews of 114 patients
with matching genotypes and known source patients identified five
patients who had only casual contact with the source patient (Cronin
2002).
The following are several possible scenarios in which two persons
who have isolates with matching genotypes might not be involved
in the same chain of recent transmission.
- The genotyping techniques used lack the power to discriminate
between two isolates that are different. For example, if two isolates
have matching PCR genotypes but different RFLP genotypes, the
two isolates would be considered genetically different. If only
the two PCR tests were conducted, however, and not the RFLP, the
two isolates would be considered matched, even though they were
genetically different.
- The transmission of common endemic strains of M. tuberculosis
occurs in relatively closed populations. In a study conducted
by the Arkansas Department of Health and CDC, experienced field
workers interviewed 78 patients whose isolates had genotypes that
matched those of at least one other person; epidemiologic links
could be identified for only 33 (42%). The authors concluded that
in a rural setting the presence of matching genotypes often is
not associated with recent transmission (Braden 1997).
- An extensive outbreak of TB in the past leads to a large number
of persons becoming latently infected with an identical strain
of M. tuberculosis. Several years later these persons may
reactivate their infections and develop active TB with matching
genotypes. A contact investigation of these persons might well
fail to establish an epidemiologic link between two patients,
since the actual link connecting these patients occurred several
years in the past. In a study by the New York City TB program
of genotyping results for persons with MDR TB, of 153 persons
who had matching genotypes, only 25 (19.8%) had epidemiologic
links identified, and most persons had been exposed to patients
diagnosed with TB before the study period, at a time when New
York City was experiencing large outbreaks of MDR TB (Munsiff
2002).
- Cultures become contaminated in the laboratory and cause a
false-positive culture to be reported. For example, during the
processing of two specimens in a clinical laboratory, if a break
in technique occurs and one specimen that contains M. tuberculosis
is allowed to contaminate another specimen that does not, subsequent
genotyping tests will report that the two specimens have isolates
of M. tuberculosis that have matching genotypes.
Other types of laboratory error can occur and result in two isolates
incorrectly being reported as having matching genotypes. For example,
a submitting laboratory can mislabel specimens or ship the wrong
specimen to the genotyping laboratory or the genotyping laboratory
can mislabel specimens, perform the genotyping incorrectly, or mix
up reports so that one specimen is reported to have the results
of another specimen.
Genotypes Do Not Match but Known Epidemiologic Links Identified
As has been stated, there are multiple reasons why matching genotypes
might not be associated with recent transmission. In contrast, it
is much less common for isolates with nonmatching genotypes to be
the result of recent transmission, since genotypes that do not match
indicate that the two isolates are genetically different (unless
there was a laboratory error in reporting those nonmatching results).
Other than laboratory error, two possibilities have been reported
where recent transmission was documented even though the isolates
had nonmatching genotypes.
The first possibility occurs when the genotype of a particular
strain of M. tuberculosis changes slightly over time. The
DNA sequences on which genotypes are based have a tendency to mutate,
change location, and be duplicated or deleted over time, forming
the basis for the diversity of the genotypes. In another large outbreak
of multidrug-resistant (MDR) TB in New York and surrounding areas,
there was strong epidemiologic evidence that over 300 patients were
infected with the same strain of M. tuberculosis (Bifani
1996). Genotyping with IS6110-based RFLP showed that the
patients were infected with isolates that had multiple closely related
genotypes. The investigators concluded that the M. tuberculosis
strain that caused this outbreak had evolved during the prolonged
outbreak. In practice, slight changes in RFLP patterns appear to
be more common than changes with the PCR-based tests. Therefore,
when a TB program requests that the genotyping laboratory run an
RFLP on isolates in a PCR cluster, the genotyping laboratory report
will flag isolates that have different but very similar RFLP patterns.
Note that RFLP patterns will not be compared among isolates that
do not belong to the same PCR cluster. Similarly, spoligotype patterns
and MIRU results can change slightly over time. In this case, closely
related isolates can show spoligotype or MIRU designations that
differ by only one character. The genotyping laboratory report will
not flag instances of PCR designations that differ slightly. However,
if two patients have known epidemiologic links but have different
PCR designations, the spoligotype and MIRU designations should be
examined to determine if they are closely related. Isolates
that match in all but one digit in the MIRU type or that have spoligotypes
that differ in one region (as a result of loss of one or more spacers)
should be considered closely related. Programs should consult
with the genotyping laboratory or CDC to determine if RFLP typing
of such isolates is warranted.
The second possibility occurs when a person is infected with more
than one strain of M. tuberculosis. If a person with a dual
infection transmits one strain to someone else, subsequent specimens
from the two persons might grow the same strain of M. tuberculosis
or might grow different strains, depending on which of the two strains
grew from the specimen from the person with the dual infection.
If different strains are grown from the two specimens, the isolates
will have nonmatching genotypes. Although this situation has been
reported, it is probably rare (Yeh 1999).
As mentioned, isolates with nonmatching genotypes are only rarely
associated with recent transmission, even if epidemiologic links
are identified. A much more common explanation of nonmatching genotypes
where epidemiologic links have been established is when the links
do not reflect recent transmission, and the presumed source case
was not the real source of transmission. A study from the National
Tuberculosis Genotyping and Surveillance Network documented common
scenarios in which a presumed epidemiologic link between a source
case and a secondary case was not confirmed by genotyping (Bennett
2002). Two common scenarios occurred when the presumed secondary
case-patient was born in a foreign country or when the presumed
source case-patient was culture-positive but sputum smear-negative.
Table 4.5. Conceptual framework for combining epidemiologic
links with genotyping results for two tuberculosis cases. Percentages
refer to the frequency that these categories were detected among
2,055 cases interviewed as part of the NTGSN study in the four
sites that conducted cluster investigations of all TB patients
with matching genotypes who did not have epidemiologic links identified
during contact investigations (unpublished NTGSN data).
- Genotypes match (27%)
- Known epidemiologic links identified (8%)
Interpretation: Two cases probably involved in the same
chain of recent transmission.
- No or only possible epidemiologic links identified
(19%)
- Cases involved in same chain of recent transmission
Interpretation: Failure to identify known epidemiologic
links despite recent transmission could have been
the result of:
- No contact investigation was conducted; OR
- Focus of contact investigation was too narrow
(e.g., focus was limited to household contacts
when transmission occurred at a leisure site);
OR
- Patients hid information about contacts;
OR
- Inadequate interviews failed to identify
epidemiologic links; OR
- Contact between cases was casual and unrecognized
by them
- Cases not involved in same chain of recent transmission
Interpretation: Matching genotypes in the absence
of recent transmission could have been the result
of:
- Transmission of an endemic strain in a relatively
closed population; OR
- False-positive culture(s); OR
- Laboratory error
- Genotypes do not match (73%)
- Known epidemiologic links identified (4%)
- Cases involved in same chain of recent transmission
Interpretation: Nonmatching genotypes in two persons
involved in the same chain of recent transmission
could be the result of:
- Genotypes that changed slightly over time;
OR
- Coinfection with >1 strain of M. tuberculosis;
OR
- Laboratory error
- Cases not involved in same chain of recent transmission
Interpretation: Nonmatching genotypes in two persons
not involved in the same chain of recent transmission
could be the result of:
- Misleading epidemiologic links identified
- No or only possible epidemiologic links identified
(69%)
Interpretation: No evidence of recent transmission
|
Nonmatching Genotypes with No or Only Possible Epidemiologic
Links Identified
When isolates from two persons have nonmatching genotypes and interviews
with the two persons reveal no known epidemiologic links, the two
are probably not involved in the same chain of recent transmission.
TB in both persons could represent reactivation of a latent TB infection,
or one person might have been recently infected from an unidentified
source case. This could happen when the source case did not live
in the same TB jurisdiction or if no isolate was available for genotyping
from the source case. No additional information needs to be gathered
to strengthen this conclusion.
Last Reviewed: 05/18/2008 Content Source: Division of Tuberculosis Elimination
National Center for HIV/AIDS, Viral Hepatitis, STD, and TB Prevention
|